Hyperspectral Sea Ice Image Classification Based on the Spectral-Spatial-Joint Feature with Deep Learning

Sea ice is one of the causes of marine disasters. The classification of sea ice images is an important part of sea ice detection. The labeled samples in hyperspectral sea ice image classification are difficult to acquire, which causes minor sample problems. In addition, most of the current sea ice c...

Full description

Bibliographic Details
Published in:Remote Sensing
Main Authors: Yanling Han, Yi Gao, Yun Zhang, Jing Wang, Shuhu Yang
Format: Article in Journal/Newspaper
Language:English
Published: MDPI AG 2019
Subjects:
Q
Online Access:https://doi.org/10.3390/rs11182170
https://doaj.org/article/a2d469cb2b6949dd9b764c9bc4530fa3
id ftdoajarticles:oai:doaj.org/article:a2d469cb2b6949dd9b764c9bc4530fa3
record_format openpolar
spelling ftdoajarticles:oai:doaj.org/article:a2d469cb2b6949dd9b764c9bc4530fa3 2023-05-15T15:35:07+02:00 Hyperspectral Sea Ice Image Classification Based on the Spectral-Spatial-Joint Feature with Deep Learning Yanling Han Yi Gao Yun Zhang Jing Wang Shuhu Yang 2019-09-01T00:00:00Z https://doi.org/10.3390/rs11182170 https://doaj.org/article/a2d469cb2b6949dd9b764c9bc4530fa3 EN eng MDPI AG https://www.mdpi.com/2072-4292/11/18/2170 https://doaj.org/toc/2072-4292 2072-4292 doi:10.3390/rs11182170 https://doaj.org/article/a2d469cb2b6949dd9b764c9bc4530fa3 Remote Sensing, Vol 11, Iss 18, p 2170 (2019) sea ice hyperspectral images (HSIs) gray-level co-occurrence matrix (GLCM) spectral-spatial-joint features unlabeled samples convolutional neural network (CNN) Science Q article 2019 ftdoajarticles https://doi.org/10.3390/rs11182170 2022-12-31T16:36:07Z Sea ice is one of the causes of marine disasters. The classification of sea ice images is an important part of sea ice detection. The labeled samples in hyperspectral sea ice image classification are difficult to acquire, which causes minor sample problems. In addition, most of the current sea ice classification methods mainly use spectral features for shallow learning, which also limits further improvement of the sea ice classification accuracy. Therefore, this paper proposes a hyperspectral sea ice image classification method based on the spectral-spatial-joint feature with deep learning. The proposed method first extracts sea ice texture information by the gray-level co-occurrence matrix (GLCM). Then, it performs dimensionality reduction and a correlation analysis of the spectral information and spatial information of the unlabeled samples, respectively. It eliminates redundant information by extracting the spectral-spatial information of the neighboring unlabeled samples of the labeled sample and integrating the information with the spectral and texture data of the labeled sample to further enhance the quality of the labeled sample. Lastly, the three-dimensional convolutional neural network (3D-CNN) model is designed to extract the deep spectral-spatial features of sea ice. The proposed method combines relevant textural features and performs spectral-spatial feature extraction based on the 3D-CNN model by using a large amount of unlabeled sample information. In order to verify the effectiveness of the proposed method, sea ice classification experiments are carried out on two hyperspectral data sets: Baffin Bay and Bohai Bay. Compared with the CNN algorithm based on a single feature (spectral or spatial) and other CNN algorithms based on spectral-spatial features, the experimental results show that the proposed method achieves better sea ice classification (98.52% and 97.91%) with small samples. Therefore, it is more suitable for classifying hyperspectral sea ice images. Article in Journal/Newspaper Baffin Bay Baffin Bay Baffin Sea ice Directory of Open Access Journals: DOAJ Articles Baffin Bay Remote Sensing 11 18 2170
institution Open Polar
collection Directory of Open Access Journals: DOAJ Articles
op_collection_id ftdoajarticles
language English
topic sea ice
hyperspectral images (HSIs)
gray-level co-occurrence matrix (GLCM)
spectral-spatial-joint features
unlabeled samples
convolutional neural network (CNN)
Science
Q
spellingShingle sea ice
hyperspectral images (HSIs)
gray-level co-occurrence matrix (GLCM)
spectral-spatial-joint features
unlabeled samples
convolutional neural network (CNN)
Science
Q
Yanling Han
Yi Gao
Yun Zhang
Jing Wang
Shuhu Yang
Hyperspectral Sea Ice Image Classification Based on the Spectral-Spatial-Joint Feature with Deep Learning
topic_facet sea ice
hyperspectral images (HSIs)
gray-level co-occurrence matrix (GLCM)
spectral-spatial-joint features
unlabeled samples
convolutional neural network (CNN)
Science
Q
description Sea ice is one of the causes of marine disasters. The classification of sea ice images is an important part of sea ice detection. The labeled samples in hyperspectral sea ice image classification are difficult to acquire, which causes minor sample problems. In addition, most of the current sea ice classification methods mainly use spectral features for shallow learning, which also limits further improvement of the sea ice classification accuracy. Therefore, this paper proposes a hyperspectral sea ice image classification method based on the spectral-spatial-joint feature with deep learning. The proposed method first extracts sea ice texture information by the gray-level co-occurrence matrix (GLCM). Then, it performs dimensionality reduction and a correlation analysis of the spectral information and spatial information of the unlabeled samples, respectively. It eliminates redundant information by extracting the spectral-spatial information of the neighboring unlabeled samples of the labeled sample and integrating the information with the spectral and texture data of the labeled sample to further enhance the quality of the labeled sample. Lastly, the three-dimensional convolutional neural network (3D-CNN) model is designed to extract the deep spectral-spatial features of sea ice. The proposed method combines relevant textural features and performs spectral-spatial feature extraction based on the 3D-CNN model by using a large amount of unlabeled sample information. In order to verify the effectiveness of the proposed method, sea ice classification experiments are carried out on two hyperspectral data sets: Baffin Bay and Bohai Bay. Compared with the CNN algorithm based on a single feature (spectral or spatial) and other CNN algorithms based on spectral-spatial features, the experimental results show that the proposed method achieves better sea ice classification (98.52% and 97.91%) with small samples. Therefore, it is more suitable for classifying hyperspectral sea ice images.
format Article in Journal/Newspaper
author Yanling Han
Yi Gao
Yun Zhang
Jing Wang
Shuhu Yang
author_facet Yanling Han
Yi Gao
Yun Zhang
Jing Wang
Shuhu Yang
author_sort Yanling Han
title Hyperspectral Sea Ice Image Classification Based on the Spectral-Spatial-Joint Feature with Deep Learning
title_short Hyperspectral Sea Ice Image Classification Based on the Spectral-Spatial-Joint Feature with Deep Learning
title_full Hyperspectral Sea Ice Image Classification Based on the Spectral-Spatial-Joint Feature with Deep Learning
title_fullStr Hyperspectral Sea Ice Image Classification Based on the Spectral-Spatial-Joint Feature with Deep Learning
title_full_unstemmed Hyperspectral Sea Ice Image Classification Based on the Spectral-Spatial-Joint Feature with Deep Learning
title_sort hyperspectral sea ice image classification based on the spectral-spatial-joint feature with deep learning
publisher MDPI AG
publishDate 2019
url https://doi.org/10.3390/rs11182170
https://doaj.org/article/a2d469cb2b6949dd9b764c9bc4530fa3
geographic Baffin Bay
geographic_facet Baffin Bay
genre Baffin Bay
Baffin Bay
Baffin
Sea ice
genre_facet Baffin Bay
Baffin Bay
Baffin
Sea ice
op_source Remote Sensing, Vol 11, Iss 18, p 2170 (2019)
op_relation https://www.mdpi.com/2072-4292/11/18/2170
https://doaj.org/toc/2072-4292
2072-4292
doi:10.3390/rs11182170
https://doaj.org/article/a2d469cb2b6949dd9b764c9bc4530fa3
op_doi https://doi.org/10.3390/rs11182170
container_title Remote Sensing
container_volume 11
container_issue 18
container_start_page 2170
_version_ 1766365415834910720